© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. MLP Post
    5. 0-MLP-SM
    6. 27712
    Prev
    Next
    INDEX
    Explanations

    instances of the word "examine" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__mlp-sm_processed/0-mlp-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.0.hook_mlp_out
    Hook Layer
    0
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    156
    +0.29
    1.7%
    376
    +0.20
    1.2%
    115
    +0.18
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    95
    +0.29
    0.01
    498
    +0.20
    0.01
    345
    +0.18
    0.01
    Negative Logits
    ı
    -1.49
    itsu
    -1.44
    umab
    -1.43
     flourish
    -1.38
    ></
    -1.33
    ksen
    -1.32
    ensen
    -1.32
     quest
    -1.31
    clud
    -1.31
    dered
    -1.31
    POSITIVE LOGITS
     him
    1.76
    iqu
    1.62
    es
    1.59
     them
    1.57
    them
    1.50
     how
    1.49
    yg
    1.44
    eing
    1.43
    a
    1.40
    ldots
    1.33
    Activations Density 0.556%

    No Known Activations